Multi-armed Bandit Problem in Alteryx

[a150138]

Hi Alteryx!

Is anyone familiar with the Multi-armed Bandit problem (see: https://www.analyticsvidhya.com/blog/2018/09/reinforcement-multi-armed-bandit-scratch-python/)?

Looking at the code, it seems that the use of iterative macro is necessary. Anyone has an idea how to proceed with this, like a sample workflow?

Thanks all.

- Tiff

[JohnJPS · Accepted solution]

Hi @a150138 

You're right that some sort of iterative macro would probably be needed, however before doing that, since the Python code is literally right there ready to cut and paste, I would consider using the Python tool, or perhaps the R if that works better for you. I've attached a workflow that duplicates their example in Alteryx with both R and Python.

Notes:

• I did rearrange things a little
• I output only the "total reward" figures that are calculated
• I commented out anything that wasn't contributing to that calculation
• Python's "Alteryx.read" didn't play nice with my laptop's crazy security, so I read the physical file from within the Python, rather than grabbing input #1... you can comment out the physical file read and comment back in the Alteryx.read(#1) and it should work for you.

Hope that helps!

John

[a150138]

Hi John,

Very useful and thanks for the guidance! I will explore the R & Python code inside and see whether I can take it a step further by converting it into a macro (Still learning).

🙂